Abstract:
Mobile services and applications are being replaced with new location-aware applications and
services. Although mobile commerce enables access to services and applications regardless of
the location of user, in many situations the specific location of the user is highly important. Map
applications, weather applications, tracking applications, commerce applications heavily use
geo-location information to improve user experience. Use of GPS data has been the trend for
location based services; nevertheless, this has been a huge over-head on users in battery life.
GPS location information requires special hardware support in the handset, which is not
available on many mobile devices. These limitations made us think about less costly location
information through GSM networks. There is a trade-off between location accuracy and cost of
the location information.In this paper, we attempt to provide a realistic assessment on use of the
mobile cell level information for identifying location and the movement of a mobile device. It
will elaborate on our dataset used for experiments; advantages of using the GSM based location
information, challenges in using GSM cell level data for location and movement prediction, and
possibilities with more accurate GSM level information.
Methodology
GSM Location information can be categorized into five levels with increasing accuracy from 1
to 5.The Sri Lankan mobile service providers mostly use pull services that request the required
user data only when necessary. Thus, it is only able to retrieve the level 1 (Cell ID) and level 2
(Cell is divided into three directions and will give the users location using one of the three
directions with the cell ID) data in the current situation.The format of the level 2 data set we
received from our service provider is as follows:
A Data line is formatted as IMSI number, date, time and CGI.
IMSI:The International Mobile Subscriber Identity is used to identify the user of a cellular
network and is a unique identification associated with all cellular networks. In the dataset, it is
anonymized by hashing the IMSI due to privacy concerns.
CGI:Cell Global Identity is a standard identifier for mobile phones cells. It geographically
connects mobile phones. If a cell phone is connected to a GSM network then the position of that
particular cell phone can be determined using CGI of the cell, which is covering that cell phone.
For the following analysis and evaluation, we have used the above dataset.